Concrete and steel bridge Structural Health Monitoring—Insight into choices for machine learning applications

Structural Health Monitoring (SHM) systems have been installed on bridges across the world at an increasing rate in recent years, providing vital data for bridge assessment and maintenance. Machine Learning (ML) is efficient in data analyses such as classification and regression, and capable of improving its accuracy by learning from data without the need for step-to-step programming. The implementation of ML methods in bridge SHM studies has become more popular in recent years for its ability to detect damages on concrete and steel caused by material deterioration and to perform condition assessment on bridge structures. There have been several review articles discussing ML applications in SHM which mostly provide broad discussions across different civil engineering structures. In this article, different ML applications in the bridge SHM study are summarised and discussed. Detailed critiques of each types of ML applications are provided. Finally, recommendations are made for the future study of ML applications in bridge SHM to fill the current research gaps.